Video systems rely greatly on the encoding process to compress video in preparation for transmission, or to transform it from one format to another. Encoding video involves techniques that seek to ensure as much compression as possible with as few errors or artifacts as possible, with as high quality as possible. One of the main causes of artifacts and low quality or quality degradation in video processing is noise. Therefore, estimating noise before encoding a video has become an essential part of the video encoding process.
Noise level estimation in the analog input TV signal during the process of video encoding allows improving of the qualitative and quantitative characteristics and parameters of the encoded video. Qualitative characteristics can be, for example, using different filters coefficients for the video preprocessing or different quantization levels for subjective increase of the video quality of the encoded data. And quantitative characteristics can be, for example, using different bit rates to conserve the space on the external storage device for the encoded data.
Noise sources act on all components of the analog TV signal during its distribution. As a result the lines from the active video region and other parts of the TV signal are exposed to noise. The lines from the active video region are the lines displayed by the TV receiver, e.g. lines 22-240 of each field of the NTSC signal. Other parts of the TV signal are lines from the vertical blanking region, e.g., lines 10-21 of each field of the NTSC signal. The latter are not included into the output video encoded data and so they are not displayed when encoded video data is decoded and presented.
Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of skill in the art, through comparison of such systems with some aspects of the present invention as set forth in the remainder of the present application with reference to the drawings.